Linux System Administration phần 5 pps

You can use Gnome RPM in a similar manner to update, delete, or query RPM packages. GUI configuration utilities in other distributions, such as Caldera and SuSE, differ in details but have similar functionality. Installing a Debian Package Debian, Xandros (formerly Corel), and Libranet Linux all use Debian packages rather than RPMs. Debian packages are incompatible with RPM packages, but the basic principles of operation are the same. Like RPMs, Debian packages include dependency information, and the Debian package utilities maintain a database of installed packages, files, and so on. You use the dpkg command to install a Debian package. This command's syntax is similar to that of rpm: dpkg [options][action] [package−files|package−name] The action is the action to be taken; common actions are summarized in Table 8.3. The options (Table 8.4) modify the behavior of the action, much like the options to rpm. Table 8.3: dpkg Primary Actions dpkg Action Description −i or −−install Installs a package. −−configure Reconfigures an installed package: Runs the post−installation script to set site−specific options. −r or −P or −−remove or −−purge Removes a package. −p or −−print−avail Displays information about a package. −l pattern or −−list pattern Lists all installed packages whose names match pattern. −L or −−listfiles Lists the installed files associated with a package. −C or −−audit Searches for partially installed packages and suggests what to do with them. Table 8.4: Options to Fine−Tune dpkg Actions dpkg Option Used with Actions Description −−root=dir All Modifies the Linux system using a root directory located at dir. Can be used to maintain one Linux installation discrete from another one, say during OS installation or emergency maintenance. −B or −−auto−deconfigure −r Disables packages that rely upon one being removed. −−force−things Assorted Forces specific actions to be taken. Consult the dpkg man page for details of things this option does. −−ignore−depends=package −i, −r Ignores dependency information for the specified package. −−no−act −i, −r Checks for dependencies, conflicts, and other problems without actually installing the package. −−recursive −i Installs all packages, matching the package name wildcard in the specified directory and all subdirectories. 189 −G −i Doesn't install the package if a newer version of the same package is already installed. −E or −−skip−same−version −i Doesn't install the package if the same version of the package is already installed. As an example, consider the following command, which installs the samba_2.0.7−3.4_i386.deb package: # dpkg −i samba_2.0.7−3.4_i386.deb If you're upgrading a package, you may need to remove an old package. To do this, use the −r option to dpkg, as in # dpkg −r samba Note It's possible to use both RPM and Debian packages on one computer, and in fact some distributions (such as Xandros/Corel Linux) explicitly support this configuration. Using both package formats reduces the benefits of both, however, because the two may introduce conflicting packages and they cannot share their dependency information. It's therefore best to use just one package format. The Debian package system includes a set of utilities known collectively as the Advanced Package Tool (APT). The most important program in this package is apt−get, which you can use to automatically or semiautomatically maintain a system. This tool is described briefly in the upcoming section, "Update Utilities." Some Debian−based Linux distributions, such as Xandros/Corel Linux, include GUI front−ends to dpkg; they're similar to the Gnome RPM program for RPM−based systems. If you're more comfortable with GUI tools than with command−line tools, you can use the GUI tools much as you'd use Gnome RPM. Installing a Tarball If you have Slackware Linux or another distribution that uses tarballs, you can install software by using the Linux tar utility. You can also use this method if you want to install a tarball on a Linux distribution that uses a package management tool. We recommend using RPM or Debian packages whenever possible, however. WarningWhen installing a program over an older version on Slackware Linux, the new files should overwrite the old ones. If you install a tarball on a system that normally uses packages, however, or if you install a tarball that was created using a different directory structure than what your current system uses, you may end up with duplicate files. This can cause confusion, because you might end up using the old binaries after installing the new ones. You should therefore remove the old package as well as you can before installing a binary tarball. Check the directory structure inside a tarball by typing tar tvfz package.tgz. This command displays all the files in the tarball, including their complete paths. Tarball installation is a fairly straightforward matter. As root, you issue commands similar to the following, which install the files from the samba.tgz file located in the /root directory: # cd / # tar xvfz /root/samba.tgz 190 Note that the first command (cd /) is important; without it, you'll install the files under the directory you're currently in, not in the usual directory tree. (It is possible, however, that the tarball might have to be installed under some directory other than /, in which case you should follow the directions that come with the package.) Note Chapter 9 describes the tar utility in greater detail. Administrator's Logbook: Binary Tarball Installation System: E12345678 Action: Installed Samba from binary tarball, samba.tgz. Files located in /opt/samba. Compiling Source Code It's frequently desirable or necessary to compile a program from source code. Situations when you might want to do this include the following: You can't find a binary package for the program. This is particularly likely when you run Linux on a non−x86 system, such as a Macintosh. • The binary packages you've found rely upon different support libraries than what you have. Recompiling often works around this problem, although sometimes the source code itself requires libraries other than what you have. • You want to enable options that are not used in the available binary packages. These options may optimize a package for your computer or add functionality. • You want to modify the source code. If there's a bug that's been fixed since the last binary package was released, or if you want to add a feature or modify a program in some way, you have little choice but to compile from source code. • In the first two cases, you can often compile from a source RPM, which is an RPM file containing source code. It's also possible to create Debian packages from source code, given appropriate control files. In the latter two cases, it's easiest to obtain a source tarball, make your modifications, and install directly from the compiled code. Creating a package from modified or optimized source code is seldom worthwhile for a one−computer installation. If you maintain several Linux computers, though, you might want to read the RPM HOWTO document or Ed Bailey's Maximum RPM to learn how to generate a binary RPM from a source code tarball. You can then install the customized package on all your computers after compiling it on just one system. Compiling from Packages Source RPM files are identified by the presence of .src. in the filename, rather than .i386. or some other architecture identifier. For example, samba−2.2.3a−6.src.rpm is the source RPM for the Samba package that comes with Red Hat 7.3; samba−2.2.3a−6.i386.rpm is the matching binary for x86 computers. If you wanted to compile the source RPM on Yellow Dog Linux (for Macintosh and other PPC−based systems), the result would be samba−2.2.3a−6.ppc.rpm. Note Some non−source RPM files are architecture independent. These can contain documentation, fonts, scripts, and so on. They're identified by a .noarch. filename component. These RPMs can be installed on systems using any CPU. 191 To compile a source RPM package, you add the −−rebuild operation to the rpm command, thus: # rpm −−rebuild samba−2.2.3a−6.src.rpm If all goes well, you'll see a series of compilation commands run as a result. These may take anywhere from a few seconds to several hours to run, depending on the package and your computer's speed. On a typical 500MHz Intel−architecture computer, most packages compile in a few minutes. Building a package requires that you have necessary support libraries installed—not just the libraries required by the final binary package, but also the matching development libraries. These libraries aren't always included in source RPM dependency information, so it's not unusual to see a compile operation fail because of a missing library. If this happens, examine the error message and then check the list of requirements on the program's home page. With luck, the failure message will bear some resemblance to a requirement listed on the package's home page. You can then locate an appropriate development RPM (which usually contains devel in its name), install it, and try again. Tip You can often use the command rpm −qpi packagefile to locate the program's home page. The package maintainer often has a home page, as well. Once a package has successfully compiled, you'll find one or more matching binary RPM files in the /usr/src directory tree. Most distributions name a directory in this tree after themselves, such as /usr/src/redhat on Red Hat systems. This directory contains an RPMS directory, which in turn has one or more subdirectories named after the architecture, such as i386 or ppc. (Most packages built on Intel−architecture computers place binaries in the i386 subdirectory, but some use i586 or some other name.) The RPM files you find in this subdirectory are binary packages that you can install just like any other binary RPM. Most source RPMs create one binary RPM file when built, but some generate multiple binary RPM files. Administrator's Logbook: RPM Source File Installation System: E12345678 Action: Compiled Samba 2.2.3a from source RPM & installed resulting binary RPM. It's possible to compile a Debian package from source code, but the process is somewhat different for this. Instead of a binary Debian package, you must locate and use a control file, which you can use in conjunction with a regular source code tarball. Compiling Tarballs If you don't want to or can't create a package file, you can compile source code from an original source tarball and install the compiled software directly. You then give up the advantages of RPM or Debian packages, however. Whenever possible, it's best to use a binary package or to create your own binary package from a source package, rather than install directly from a source tarball. Note Some administrators prefer using original source tarballs because they know the source code hasn't been modified by the package maintainer, as is quite common with RPM (including source RPM) files. 192 You can unpack a tarball using a command like tar xvzf sourcecode.tgz. This usually produces a subdirectory containing the source code distribution. You can unpack this tarball in a convenient location in your home directory, in the /root directory, in the /usr/src directory, or somewhere else. Some operations involved in compiling and installing the code may require root privileges, though, so you might not want to use your home directory. Unfortunately, it's impossible to provide a single procedure that's both complete and accurate for all source code tarballs. This is because no two source code packages are exactly alike; each developer has his or her own style and preferences in compilation and installation procedures. Some elements are quite commonly included, however: Documentation Most source tarballs have one or more documentation files. Sometimes these appear in a subdirectory called doc or documentation. Other times there's a README or INSTALL file, or OS−specific files (README.linux, for instance). Read the ones that are appropriate. Configuration options Most large programs are complex enough that they require precompilation configuration for your OS or architecture. This is often handled through a script called configure. The script checks for the presence of critical libraries, compiler quirks, and so on, and creates a file called Makefile that will ultimately control compilation. A few programs accomplish the same goal through some other means, such as typing make config. Sometimes you must answer questions or pass additional parameters to a configuration script. Compilation To compile a package, you must usually type make. For some packages, you must issue individual make commands for each of several subcomponents, as in make main. The compilation process can take anywhere from a few seconds to several hours, depending on the package and your computer's speed. Installation Small packages sometimes rely on you to do the installation; you must copy the compiled binary files to /usr/local/bin or some other convenient location. You may also need to copy man page files, configuration files, and so on. The package's documentation will include the details you need. Other packages have a script called install or a make option (usually typing make install) to do the job. Post−installation configuration After installing the software, you may need to configure it for your system by editing configuration files. These may be located in users' home directories, in /etc, or elsewhere. The program's documentation should provide details. The traditional location for packages compiled locally is in the /usr/local directory tree—/usr/local/bin for binaries, /usr/local/man for man pages, and so on. This placement ensures that installing a program from a source tarball won't interfere with package−based programs, which typically go elsewhere in the /usr tree. Most source tarballs include default installation scripts that place their contents in /usr/local, but a few don't follow this convention. Check the program's documentation to find out where it installs. Administrator's Logbook: Source Code Package Installation System: E12345678 193 Action: Compiled & installed Samba 2.2.3a. Located in /usr/local/samba. Kernel Compilation The Linux kernel is a particularly critical and complex component on any Linux system. It therefore deserves special consideration in any discussion of software installation and maintenance. Although you can install a precompiled updated kernel much as you can other precompiled packages, doing your own kernel compilation offers certain advantages, as described shortly. The kernel compilation and installation process has its own quirks, so this section covers the process in detail, starting at setting the compilation options and proceeding through rebooting the computer to use the new kernel. Why Compile Your Kernel? With any luck, your computer booted and ran immediately after you installed Linux on it. This fact means that the Linux kernel provided with the distribution works. Why, then, should you go to the bother of compiling a new kernel? The most important advantages to custom kernel compilation are: Architecture optimization The kernel includes optimization options for each of several classes of CPU—80386, 80486, and so on. Most distributions ship with kernels that are optimized for 80386 CPUs. By compiling a kernel for your particular CPU model, you can squeeze a little extra speed out of your system. Removing unnecessary drivers The default kernel includes drivers for a wide variety of hardware components. In most cases, these drivers do no harm because they're compiled as modules (separate driver files), which aren't loaded unless necessary. A few are compiled into the kernel proper, however. These consume memory unnecessarily, thus degrading system performance slightly. Adding drivers You may need to add a new or experimental driver to your system. This may be necessary if you're using an unusually new component, or if there's a bug fix that's not yet been integrated into the main kernel tree. Such changes often require you to patch the kernel—to replace one or more kernel source code files. For details on how to do this, check with the site that provides the new driver. Changing options You may want to change options related to drivers, in order to optimize performance or improve reliability. As you examine the kernel configuration procedure, you'll see many examples of such options. Upgrading the kernel You may want to run the latest version of the kernel. Sometimes you can obtain an upgrade in precompiled form, but occasionally you'll have to compile a kernel from source code. Of course, kernel compilation isn't without its drawbacks. It takes time to configure and compile a kernel. It's also possible that the kernel you compile won't work. Be sure to leave yourself a way to boot using the old kernel, or you'll have a hard time booting your system after a failed upgrade. (This chapter describes how to boot the computer into either the old or the new kernel.) 194 On the whole, compiling your own kernel is something that every Linux system administrator should be able to do, even if it's not something you do on every system you maintain. Using a custom−compiled kernel helps you optimize your system and use cutting−edge drivers, which can give your system an advantage. In some cases, this is the only way you can get certain features to work (as with drivers for particularly new hardware). Obtaining a Kernel Before you can compile a kernel, you must obtain one. As for other software packages, you obtain a kernel either precompiled or in source code form; and in RPM, Debian package, or tarball form. We favor installing a kernel from source tarball form, because it allows you to be sure you're working from an original standard base. Kernel RPMs, in particular, are often modified in various ways. Although these modifications can sometimes be useful, they can also interfere with the smooth installation of patches should they be needed. (On the other hand, the kernels distributed as RPMs sometimes include the very patches you might want to install, thus simplifying matters.) One of the best places to look for a kernel is http://www.kernel.org/. This site includes links to "official" kernel source tarballs. You can also find kernel files on major FTP sites, such as ftp://ibiblio.org/. If you want to use an RPM or Debian package, check for kernel source code files from your distribution's maintainer. If you use an RPM or Debian kernel package, you may need to download two files: one with the kernel source code proper, and one with the kernel header files. Tarballs typically include both sets of files in a single tarball. A complete 2.4.18 kernel tarball is 29MB in size; a 2.5.5 kernel tarball is 33MB. (Kernels are also available in bzipped tar files, which are somewhat smaller than the traditional gzipped tar files. You use bzip2 to uncompress these files rather than gzip.) Because of their large size, these kernel files may take quite some time to download. Once you've downloaded the kernel tarball, you can unpack it in the /usr/src directory. The tarball creates or installs to a directory called linux. Warning If /usr/src already has a directory called linux, you should rename it to something else and create a new linux directory for the new source package. This will prevent problems caused by unpacking a new source tree over an old one, which can create inconsistencies that cause compilation failures. You can unpack a kernel tarball just as you would any other source code tarball: # tar xvzf ~/linux−2.4.18.tar.gz If your source tarball uses bzip2 compression, you can use a command similar to the following to extract it: # tar xvf ~/linux−2.4.18.tar.bz2 −−use−compress−program bzip2 Kernel Configuration Options Once you've extracted your kernel tarball, you can proceed to configure it. Use any of the following three commands to accomplish this task: make config This command runs a text−based configuration tool that asks you specific questions about each and every configuration option. You can't skip around 195 arbitrarily from one option to another, so this method is quite awkward. make menuconfig Like make config, this option presents a text−based configuration tool. The make menuconfig tool uses text−mode menus, though, so you can skip from one option to another. This is a good way to configure the kernel if you're using a text−based console login. make xconfig This command also uses menus for configuration, but the menus are X−based, so you can configure the kernel using mouse clicks in X. The kernel configuration options are arranged in groups. If you use make menuconfig or make xconfig, you can select one group to see a list of items in that group, as shown in Figure 8.2. Groups often have subgroups, so you may need to examine quite a few menus before you find a particular driver or option. Figure 8.2: Kernel compilation options are arranged hierarchically, with each main−menu option generating its own menu, which is displayed in a separate window when make xconfig is used. Note A new kernel configuration tool, CML2 (http://tuxedo.org/~esr/cml2/) is under development and should be integrated into the Linux kernel as part of the 2.5.x kernel series. This may change many details of kernel configuration, but the basic principles should remain unchanged. Describing every available kernel configuration option would be quite tedious, as well as inevitably incomplete, because options are constantly being added and changed. Therefore, Table 8.5 merely presents an overview of the main kernel headings in the 2.4.x kernel series. The 2.5.x kernel series is new enough at the time of this writing that its contents are almost certain to change in the near future. Kernel Version Numbers Each Linux kernel has a version number of the form x.y.z. The x number is the major version number, and in 2002 this number is 2.• 196 The y number denotes an important change to the kernel and has a special meaning. Even−numbered y values are considered stable—they're unlikely to contain major bugs, and they don't change much from one minor release to another. An odd y number denotes a development kernel, which contains features that are experimental. Development kernels may be unstable and may change substantially over time. Unless you're desperate to use a feature introduced in a development kernel, you shouldn't use one of these. • The z number represents a minor change within a given stable or development kernel. In stable kernels, these represent minor bug fixes and occasionally the addition of important new (but well−tested) drivers. Within development kernels, incrementing z numbers represent major bug fixes, added features, changes, and (being realistic) bug introductions. • When Linus Torvalds believes that a development kernel is becoming stable and contains the features he wants in that kernel, he calls a code freeze, after which point only bug fixes are added. When the kernel stabilizes enough, a new stable release is made based on the last development kernel in a series (a number of test releases may exist leading to this new stable release). At this writing, the current stable kernel version is 2.4.18, and the latest development kernel is 2.5.5. This development series will eventually lead to the release of a 2.6.0 or 3.0.0 kernel. Table 8.5: Linux 2.4.x Kernel Configuration Options Kernel Configuration Menu Item Subsumed Options Code Maturity Level Options This menu provides options allowing you to select experimental drivers and features. Loadable Module Support Modern kernels typically include many features in loadable modules (separate driver files). This menu lets you enable support for these modules and set a couple of options related to it. Processor Type and Features You can configure the system to optimize the kernel for your particular CPU, as well as enable CPU−related options such as floating−point emulation (which is not required for modern CPUs). General Setup This menu contains an assortment of miscellaneous options that don't fit anywhere else, such as types of binary program files supported by the kernel and power management features. Memory Technology Devices (MTD) This menu allows you to enable support for certain types of specialized memory storage devices, such as flash ROMs. Chances are you don't need this support on a workstation or server. Parallel Port Support Here you can add support for parallel−port hardware (typically used for printers and occasionally for scanners, removable disk drives, and other devices). Support for specific devices must be added in various other menus. Plug and Play Configuration The 2.4.x kernel includes support for ISA plug−and−play (PnP) cards. Prior kernels relied upon an external utility, isapnp, to configure these cards. You can use the kernel support or the old isapnp utility, whichever you prefer. Block Devices Block devices are devices such as hard disks whose contents are read in blocks of multiple bytes. This menu controls floppy disks, parallel−port−based removable disks, and a few other block devices. Some block devices, including most hard disks, are covered in other 197 menus. Multi−Device Support Logical Volume Management (LVM) and Redundant Arrays of Independent Disks (RAID) are advanced disk management techniques that can simplify partition resizing, increase disk performance, or improve disk reliability. The configuration of these options is beyond the scope of this book, but they can be enabled from this kernel configuration menu. Networking Options You can configure an array of TCP/IP networking options from this menu, as well as enable other networking stacks, such as DDP (used for AppleTalk networks) and IPX (used for Novell networks). Network hardware is configured in another menu. Telephony Support This menu lets you configure specialized hardware for using the Internet as a means of linking telephones. ATA/IDE/MFM/RLL Support Most x86 computers today use EIDE hard disks, and you enable drivers for these devices from this menu. Related older disk drivers are also enabled from this menu, as are drivers for EIDE CD−ROMs, tape drives, and so on. Kernel Configuration Menu Item Subsumed Options SCSI Support Here you enable support for SCSI host adapters and specific SCSI devices (disks, CD−ROM drives, and so on). Fusion MPT device support The Fusion MPT device is a unique mix of SCSI, IEEE 1394, and Ethernet hardware. You activate support for it from this menu. IEEE 1394 (FireWire) Support This menu allows you to enable support for the new IEEE 1394 (a.k.a. FireWire) interface protocol, which is used for some video and disk devices. I2O Device Support This menu allows you to use I2O devices. Intelligent Input/Output (I2O) is a new scheme that allows device drivers to be broken into OS−specific and device−specific parts. Network Device Support This menu contains options for enabling support of specific network hardware devices. This includes PPP, which is used for dial−up Internet connections. Amateur Radio Support You can connect multiple computers via special radio devices, some of which are supported by Linux through drivers in this menu. IrDA (Infrared) Support Linux supports some infrared communications protocols, which are often used by notebook and handheld computers. You can enable these protocols and hardware in this menu. ISDN Subsystem Integrated Services Digital Network (ISDN) is a method of communicating at up to 128Kbps over telephone lines. You can enable support for ISDN cards in this menu. Old CD−ROM Drivers (not SCSI, not IDE) Some old CD−ROM devices used proprietary interfaces. Linux supports these cards, but you must enable appropriate support with the settings in this menu. If you use a modern EIDE or SCSI CD−ROM, you do not need to enable any of these options. Input Core Support If you want to use a USB keyboard or mouse, enable support for these devices in this menu. You can also set a few other input device options here. Character Devices Character devices, in contrast to block devices, allow input/output one byte (character) at a time. Enable support for such devices 198 [...]... ' ' 354 0' Processor 1,3,0 103) * 1,4,0 104) 'CONNER ' 'CTT8000−S ' '1.17' Removable Tape 1 ,5, 0 1 05) 'SONY ' 'CD−R CRX140S ' '4. 05' Removable 1,6,0 106) * 1,7,0 107) * To obtain information about one specific device, use the following: cdrecord dev=1 ,5, 0 −checkdrive [root@opus root]# cdrecord dev=1 ,5, 0 −checkdrive Cdrecord 1.10 (i686−pc linux gnu) Copyright (C) 19 95 2001 Jörg Schilling scsidev: '1 ,5, 0'... output looks like this: Cdrecord 1.10 (i686−pc linux gnu) Copyright (C) 19 95 2001 Jörg Schilling Linux sg driver version: 3.1.19 Using libscg version 'schily−0 .5' scsibus0: cdrecord: Warning: controller returns wrong size for CD capabilities page 0,0,0 0) 'ATAPI−CD' 'ROM−DRIVE 50 MAX ' '50 LT' Removable CD−ROM 0,1,0 1) * 0,2,0 2) * 0,3,0 3) * 0,4,0 4) * 0 ,5, 0 5) * 0,6,0 6) * 0,7,0 7) * scsibus1: 1,0,0 100)... and the subset of the filesystem to act upon Note Because dump operates on a lower level of filesystem access than do most other backup utilities, any given dump program must have explicit support for any filesystem you use The standard Linux dump utility handles ext2/ext3 filesystems, but not FAT filesystems or the Reiser journaling filesystem Trying to use dump on such filesystems won't cause problems,... stores specific information about the tape, increasing the speed of reads Uncompressed capacities for AIT tapes are 25GB, 35GB, 50 GB, and 100GB The cost for these tapes begins at about $50 for the 25GB variety The drives start at a little over $400 Robotic Systems The emerging robotic systems vary widely in storage capacity and degree of automation At the low end are auto−loading tape libraries that... in the computer, check this menu for drivers File Systems This menu has options for supporting specific filesystems such as Linux' s native ext2fs or Windows's FAT Console Drivers In this menu you can set options relating to how Linux handles its basic text−mode display Sound You can configure your sound card drivers in this menu USB Support If your system uses any USB devices, you can enable support... compromised systems Of course, today's Linux distributions don't ship with that compromised version of named; their packages have been updated to fix the bug The point is that one must protect against bugs in important programs that can open holes in a system' s security A security problem might be discovered tomorrow in a server you run today If so, your system can be compromised Indeed, if your system. .. 8.6: URLs for Major Linux Distribution Updates Distribution Caldera Debian Libranet Mandrake Red Hat Slackware SuSE TurboLinux Xandros/Corel Yellow Dog Update URL http://www.calderasystems.com/support/security/ and ftp://ftp.caldera.com/pub/updates http://www.debian.org/security/ http://www.libranet.com/support.html http://www .linux mandrake.com/en/security/ http://www.redhat.com/apps/support/errata/... any Linux system administrator Most Linux systems today use the RPM or Debian package formats, both of which allow for easy package handling by maintaining a database of packages and the individual files associated with these packages When necessary, you can build a package file from source code, or install software without using an RPM or Debian package This approach is particularly useful for the Linux. .. secure.) Administrator's Logbook: Updating Programs System: E123 456 7 Action: Updated samba−2.0.3 to samba−2.2.3a to provide support for Windows 2000 clients Update Utilities Linux distributions are increasingly shipping with utilities designed to help you automatically or semiautomatically update your software These programs can help you keep your system up to date with minimal fuss, but they're not... main contrib non−free 3 Type apt−get update to have the system retrieve an up−to−date package list 4 Type apt−get −s −u upgrade to obtain a report on packages for which updates are available 5 Type apt−get upgrade to obtain and install all the packages found in step 4 Warning Step 5 is potentially risky, because it effectively gives control of your system to whomever maintains the package update database . Installation System: E123 456 78 193 Action: Compiled & installed Samba 2.2.3a. Located in /usr/local/samba. Kernel Compilation The Linux kernel is a particularly critical and complex component on any Linux. 2.4.18, and the latest development kernel is 2 .5. 5. This development series will eventually lead to the release of a 2.6.0 or 3.0.0 kernel. Table 8 .5: Linux 2.4.x Kernel Configuration Options Kernel. drivers. File Systems This menu has options for supporting specific filesystems such as Linux& apos;s native ext2fs or Windows's FAT. Console Drivers In this menu you can set options relating to how Linux